The present invention relates to code division multiple access communication systems based on spread sequences designed by using complete factorization of a polyphase matrix of a perfect construction (PR) filter bank. It is especially suitable for use in mobile communication system where the channels between transmitters and receivers are characterized by multipath fading. This spread sequence is applicable in mobile stations or in base stations where code division multiple access (CDMA) is utilized. In addition to mobile communication systems, this spread sequence could be utilized in personal communication system (PCS) and digital cellular system. The invention facilitates better communication over channels characterized by fading and jamming.
In mobile communication systems, the channels between transmitters and receivers are characterized by multipath fading as well as by additive noise. Various approaches have been taken to minimize these errors by means of coding and decoding in the transmitters and receivers respectively. In an alternative approach, spread sequences which are significantly longer than the data symbol have been used to minimize the errors due to multipath fading. By employing spread sequences which are significantly longer than the data symbol interval, the transmission of each data symbol of each user is spread over a wide temporal and spectral content. The data symbols spread by these sequences are overlapped with neighboring data symbols to maintain the same bandwidth. A brief description appears on xe2x80x9cSpread-Signature CDMA: Efficient multiuser communications in the presence of fadingxe2x80x9d by G. W. Wornell in IEEE Transactions on Information Theory, pp 1418-1438, Vol 41, No.5, September 1995 (U.S. Pat. No.: 5,570,351). Wornell used a structure to design longer spread sequences with binary values. Although the codes are binary valued, the transmitted streams generated via modulation are not binary values. Modulated data streams are integer valued. At the same time, the length of the code increases exponentially with the number of users. For example, if the number of users is M, overlapping is constrained to Mk, k=1,2,3, . . . i.e., the sequence length is constrained to M2,M3,M4, . . . etc.
The present invention provides a spread sequence design method for code division multiple access communication system using complete factorization of the polyphase matrix of a perfect reconstruction (PR) filter bank. The length of the spread sequence is longer than the period of the data symbol i.e., the data symbols are spread by the spread sequence and overlapped with the neighboring data symbols. A similar spread sequence is employed at the receiver to demodulate the received data stream. Multiple units could be employed at the receiver to process the delayed versions of the received signal and combine by maximal ratio, equal gain or other alternative criteria to generate decision variable. According to the present invention, the spread sequence design method includes an orthogonal code generator to generate an orthogonal matrix and cascading stages which comprise additions, subtractions and shifting operations. The number of cascading stages is determined according to the number of sequences and the length of the sequences required. Random permutation matrices are used at every cascading stage to randomize the spread sequences.
According to another aspect of the present invention, the present sequence is implemented in QPSK. This includes an orthogonal code generator to generate a Walsh-Hadamard matrix and cascading stages which comprise additions, subtractions and shifting operations. The number of cascading stages is determined according to the number of sequences and the length of the sequences required. Random permutation matrices are used at every cascading stage to randomize the spread codes. An interleaving coder encodes the input data symbols into two data series and assigning then to the I-phase and the Q-phase respectively, in a QPSK modulator. A first upsampler and sequence modulator for the data series is assigned to the I-phase. A second upsampler and sequence modulator for the data series is assigned to the Q-phase. According to another aspect of the present invention, the integer valued spread sequence design method includes a Walsh-Hadamard code generator to generate a Walsh-Hadamard matrix and cascading stages which comprise additions, subtractions and shifting operations. The number of cascading stages is determined according to the number of sequences and the length of the sequences required. Random permutation matrices are used at every cascading stage to randomize the spread sequences.
According to another aspect of the present invention, the present integer valued sequence is implemented in QPSK. This includes a Walsh-Hadamard code generator to generate a Walsh-Hadamard matrix and cascading stages which comprise additions, subtractions and shifting operations. The number of cascading stages is determined according to the number of sequences and the length of the sequences required. Random permutation matrices are used at every cascading stage to randomize the spread codes. An interleaving coder encodes the input data symbols into two data series and assigns them to the I-phase and the Q-phase respectively, in a QPSK modulator. A first upsampler and sequence modulator for the data series is assigned to the I-phase. A second upsampler and sequence modulator for the data series is assigned to the Q-phase.
Moreover, in the present invention, the number of cascading stages and permutation matrices are selected according to the number of sequences and the length of the sequences required. The length of the sequence is determined according to the performance requirement. By employing spread sequences which are longer than the period of data symbol, the transmission of each data symbol of each user is spreaded over a wide temporal and spectral extent. Furthermore, when the sequences are designed from a Walsh-Hadamard matrix and with some cascading stages, it can be implemented using simple integer arithmetic. The number of integer values level could be limited by employing proper permutation matrices. By incorporating a predetermined random permutation matrix, the sequence can be designed to be time varying sequence. Furthermore, This spread sequences are highly secured since random permutation matrices are used.
As will be appreciated, the invention is capable of other and different embodiments, as its several details are capable of modifications in various respects, all without departing from the invention. Accordingly, the drawings and description of the preferred embodiments are to be regarded as illustrative in nature and not restrictive.